Descriptives across the categories (Political_True and _False)
df_political %>% group_by(Category, measurement) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
## # A tibble: 6 x 9
## # Groups: Category [2]
## Category measurement mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Political_Fake Acc_Abs 57.7 13.4 140 1.13 58.8 21.9 85.4
## 2 Political_Fake Acc 2.96 0.536 140 0.0453 2.90 1.8 4.42
## 3 Political_Fake Fam 2.03 0.305 140 0.0258 2.02 1.29 3.51
## 4 Political_True Acc_Abs 63.1 12.5 152 1.01 64.9 25.7 88.2
## 5 Political_True Acc 3.79 0.505 152 0.0410 3.82 2.14 4.82
## 6 Political_True Fam 2.42 0.414 152 0.0336 2.33 1.7 3.61
# Basic histogram Absolute Accuracy
df_political %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "For all values collapsed across political leaning and true vs false news") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic boxplot Absolute Accuracy
df_political %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_boxplot() + theme_apa() + labs(title = "Accuracy Boxplot - Percentage", x = "Percentage Values", y = "Count", subtitle = "For all values collapsed across political leaning and true vs false news")

# Basic histogram Absolute Accuracy
df_political %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "Grouped by true vs false news, collapsed acorss political leaning") + facet_wrap(~Category)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic boxplot Absolute Accuracy
df_political %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_boxplot() + theme_apa() + labs(title = "Accuracy Boxplot - Percentage", x = "Percentage Values", y = "Count", subtitle = "For all values collapsed across political leaning and true vs false news") + facet_wrap(~Category)

# Basic histogram Relative Accuracy
df_political %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For all values collapsed across political leaning and true vs false news \n 1: extremely unlikely; 6 extremely likely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Relative Accuracy
df_political %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "Collapsed acorss political leaning \n 1: extremely unlikely; 6 extremely likely") + facet_wrap(~Category)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
df_political %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For all values collapsed across political leaning and true vs false news \n 1: not at all; 6 extremely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
df_political %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "Collapsed acorss political leaning \n 1: not at all; 6 extremely") + facet_wrap(~Category)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Descriptives across the Political_True only
df_political_true %>% group_by(political_leaning, measurement) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
## # A tibble: 6 x 9
## # Groups: political_leaning [2]
## political_leani… measurement mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Acc_Abs 66.2 11.6 76 1.33 68.5 35.6 88.2
## 2 Democrat Acc 3.94 0.459 76 0.0526 3.98 2.71 4.82
## 3 Democrat Fam 2.45 0.458 76 0.0526 2.3 1.77 3.61
## 4 Republican Acc_Abs 60.0 12.6 76 1.45 58.2 25.7 87.8
## 5 Republican Acc 3.65 0.512 76 0.0587 3.61 2.14 4.8
## 6 Republican Fam 2.39 0.364 76 0.0418 2.34 1.7 3.44
# Basic histogram Absolute Accuracy
df_political_true %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "For Political_True only, collapsed across political leaning") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Absolute Accuracy
df_political_true %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "For Political_True only") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Relative Accuracy
df_political_true %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_True only, collapsed across political leaning \n 1: extremely unlikely; 6 extremely likely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Relative Accuracy
df_political_true %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_True only \n 1: extremely unlikely; 6 extremely likely") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
df_political_true %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_True only, collapsed across political leaning \n 1: not at all; 6 extremely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
df_political_true %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_True only \n 1: not at all; 6 extremely") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Descriptives across the Political_False only
df_political_false %>% group_by(political_leaning, measurement) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
## # A tibble: 6 x 9
## # Groups: political_leaning [2]
## political_leani… measurement mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Acc_Abs 58.8 12.7 70 1.52 59.5 26.7 85.4
## 2 Democrat Acc 2.94 0.515 70 0.0615 2.92 1.8 4.4
## 3 Democrat Fam 2.05 0.322 70 0.0385 2.02 1.43 3.51
## 4 Republican Acc_Abs 56.7 14.1 70 1.68 58.7 21.9 78.9
## 5 Republican Acc 2.98 0.560 70 0.0670 2.90 1.94 4.42
## 6 Republican Fam 2.01 0.289 70 0.0345 1.96 1.29 2.65
# Basic histogram Absolute Accuracy
df_political_false %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "For Political_False only, collapsed across political leaning") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Absolute Accuracy
df_political_false %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "For Political_False only") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Relative Accuracy
df_political_false %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_False only, collapsed across political leaning \n 1: extremely unlikely; 6 extremely likely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Relative Accuracy
df_political_false %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_False only \n 1: extremely unlikely; 6 extremely likely") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
df_political_false %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_False only, collapsed across political leaning \n 1: not at all; 6 extremely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
df_political_false %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_False only \n 1: not at all; 6 extremely") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Clean Descriptives across the categories (Political_True and _False)
c_df_political %>% group_by(Category, measurement) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
## # A tibble: 6 x 9
## # Groups: Category [2]
## Category measurement mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Political_Fake Acc_Abs 59.5 10.8 118 0.991 59.4 38.2 78.3
## 2 Political_Fake Acc 2.90 0.440 118 0.0405 2.88 1.94 3.88
## 3 Political_Fake Fam 2.02 0.271 118 0.0249 2.02 1.29 2.75
## 4 Political_True Acc_Abs 63.9 10.2 128 0.904 65.2 42.9 82.9
## 5 Political_True Acc 3.83 0.427 128 0.0377 3.84 2.88 4.82
## 6 Political_True Fam 2.45 0.394 128 0.0348 2.35 1.7 3.57
# Basic histogram Absolute Accuracy
c_df_political %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "For all values collapsed across political leaning and true vs false news") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic boxplot Absolute Accuracy
c_df_political %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_boxplot() + theme_apa() + labs(title = "Accuracy Boxplot - Percentage", x = "Percentage Values", y = "Count", subtitle = "For all values collapsed across political leaning and true vs false news")

# Basic histogram Absolute Accuracy
c_df_political %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "Grouped by true vs false news, collapsed acorss political leaning") + facet_wrap(~Category)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic boxplot Absolute Accuracy
c_df_political %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_boxplot() + theme_apa() + labs(title = "Accuracy Boxplot - Percentage", x = "Percentage Values", y = "Count", subtitle = "For all values collapsed across political leaning and true vs false news") + facet_wrap(~Category)

# Basic histogram Relative Accuracy
c_df_political %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For all values collapsed across political leaning and true vs false news \n 1: extremely unlikely; 6 extremely likely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Relative Accuracy
c_df_political %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "Collapsed acorss political leaning \n 1: extremely unlikely; 6 extremely likely") + facet_wrap(~Category)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
c_df_political %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For all values collapsed across political leaning and true vs false news \n 1: not at all; 6 extremely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
c_df_political %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "Collapsed acorss political leaning \n 1: not at all; 6 extremely") + facet_wrap(~Category)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Clean Descriptives across the Political_True only
c_df_political_true %>% group_by(political_leaning, measurement) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
## # A tibble: 6 x 9
## # Groups: political_leaning [2]
## political_leani… measurement mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Acc_Abs 66.5 10.0 64 1.26 68.8 44.6 81.7
## 2 Democrat Acc 3.96 0.402 64 0.0502 3.99 3.06 4.82
## 3 Democrat Fam 2.47 0.437 64 0.0547 2.32 1.8 3.57
## 4 Republican Acc_Abs 61.3 9.79 64 1.22 58.6 42.9 82.9
## 5 Republican Acc 3.70 0.414 64 0.0518 3.62 2.88 4.73
## 6 Republican Fam 2.42 0.347 64 0.0434 2.37 1.7 3.44
# Basic histogram Absolute Accuracy
c_df_political_true %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "For Political_True only, collapsed across political leaning") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Absolute Accuracy
c_df_political_true %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "For Political_True only") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Relative Accuracy
c_df_political_true %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_True only, collapsed across political leaning \n 1: extremely unlikely; 6 extremely likely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Relative Accuracy
c_df_political_true %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_True only \n 1: extremely unlikely; 6 extremely likely") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
c_df_political_true %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_True only, collapsed across political leaning \n 1: not at all; 6 extremely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
c_df_political_true %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_True only \n 1: not at all; 6 extremely") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Clean Descriptives across the Political_False only
c_df_political_false %>% group_by(political_leaning, measurement) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
## # A tibble: 6 x 9
## # Groups: political_leaning [2]
## political_leani… measurement mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Acc_Abs 59.8 10.7 59 1.39 60.5 39.6 78.3
## 2 Democrat Acc 2.91 0.439 59 0.0571 2.91 2.15 3.75
## 3 Democrat Fam 2.03 0.271 59 0.0352 2.02 1.43 2.75
## 4 Republican Acc_Abs 59.2 10.9 59 1.42 59.4 38.2 78.1
## 5 Republican Acc 2.89 0.445 59 0.0579 2.86 1.94 3.88
## 6 Republican Fam 2.00 0.272 59 0.0354 1.93 1.29 2.59
# Basic histogram Absolute Accuracy
c_df_political_false %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "For Political_False only, collapsed across political leaning") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Absolute Accuracy
c_df_political_false %>% filter(measurement == "Acc_Abs") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Percentage", x = "Percentage Values", y = "Count", subtitle = "For Political_False only") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Relative Accuracy
c_df_political_false %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_False only, collapsed across political leaning \n 1: extremely unlikely; 6 extremely likely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Relative Accuracy
c_df_political_false %>% filter(measurement == "Acc") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Accuracy Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_False only \n 1: extremely unlikely; 6 extremely likely") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
c_df_political_false %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_False only, collapsed across political leaning \n 1: not at all; 6 extremely") + geom_vline(aes(xintercept = mean(value)), linetype="dashed")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram Familiarity
c_df_political_false %>% filter(measurement == "Fam") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Familiarity Histogram - Likert means", x = "Likert Values", y = "Count", subtitle = "For Political_False only \n 1: not at all; 6 extremely") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Item selection (politically true only)
all items, no familiarity thershold (political_true only)
#all familiarity
c_df_political_true %>% filter(measurement == "Fam") %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 4)) + labs(title = "All items, all familiarity", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

accuracy_all_fam <- c_df_political_true %>% filter(measurement == "Acc_Abs") %>% group_by(political_leaning, Par_Combined) %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_fam
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 68.7 8.84 45 1.32 70.2 45.1 81.7
## 2 Democrat Rep Favoured 61.3 11.0 19 2.53 63.5 44.6 77.6
## 3 Republican Dem Favoured 58.8 8.92 45 1.33 56.8 42.9 78.0
## 4 Republican Rep Favoured 67.3 9.35 19 2.14 66.0 51.4 82.9
#making table
accuracy_all_fam_table <- accuracy_all_fam[1:2,-1]
accuracy_all_fam %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

all items, familiarity thershold < 3 (political_true only)
# limited to three
c_df_political_true %>% filter(measurement == "Fam", value < 3) %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 4)) + labs(title = "All items, familiarity < 3", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

# need to find item numbers of those with Fam < 3 (across both dem and rep averaged)
# collapse across dem and rep - find average fam score, then exclude those that you don't need - make index from that.
accuracy_all_three_index <- c_df_political_true %>% group_by(`Item #`, `Image Name`, Par_Combined) %>% filter(measurement == "Fam") %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_three_index <- accuracy_all_three_index %>% filter(mean < 3)
accuracy_all_three_index <- accuracy_all_three_index[[1]]
# accuracy_all_three_index
# length(accuracy_all_three_index)
#given the index, now subset based on the index numbers.
accuracy_all_three <- c_df_political_true %>% filter(measurement == "Acc_Abs", `Item #` %in% accuracy_all_three_index) %>% group_by(political_leaning, Par_Combined) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_three
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 68.5 8.72 44 1.31 69.9 45.1 81.7
## 2 Democrat Rep Favoured 60.6 11.5 17 2.78 61.0 44.6 77.6
## 3 Republican Dem Favoured 58.8 9.02 44 1.36 56.8 42.9 78.0
## 4 Republican Rep Favoured 65.5 8.12 17 1.97 65.9 51.4 79.1
#making table
accuracy_all_three_table <- accuracy_all_three[1:2,-1]
accuracy_all_three %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

all items, familiarity thershold < 2.5 (political_true only)
# limited to 2.5
c_df_political_true %>% filter(measurement == "Fam", value < 2.5) %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 5)) + labs(title = "All items, familiarity < 2.5", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

# need to find item numbers of those with Fam < 2.5 (across both dem and rep averaged)
# collapse across dem and rep - find average fam score, then exclude those that you don't need - make index from that.
accuracy_all_two.five <- c_df_political_true %>% group_by(`Item #`, `Image Name`, Par_Combined) %>% filter(measurement == "Fam") %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.five <- accuracy_all_two.five %>% filter(mean < 2.5)
accuracy_all_two.five <- accuracy_all_two.five[[1]]
# accuracy_all_two.five
# length(accuracy_all_two.five)
#given the index, now subset based on the index numbers.
accuracy_all_two.five <- c_df_political_true %>% filter(measurement == "Acc_Abs", `Item #` %in% accuracy_all_two.five) %>% group_by(political_leaning, Par_Combined) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.five
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 66.4 8.89 30 1.62 68.8 45.1 81.7
## 2 Democrat Rep Favoured 56.9 11.1 11 3.34 58.7 44.6 77.6
## 3 Republican Dem Favoured 56.8 7.75 30 1.42 56.0 42.9 78.0
## 4 Republican Rep Favoured 63.1 8.15 11 2.46 65.8 51.4 76.3
#making table
accuracy_all_two.five_table <- accuracy_all_two.five[1:2,-1]
accuracy_all_two.five %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

all items, familiarity thershold < 2.4 (political_true only)
# limited to 2.4
c_df_political_true %>% filter(measurement == "Fam" , value < 2.4) %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 5)) + labs(title = "All items, familiarity < 2.4", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

# need to find item numbers of those with Fam < 2.4 (across both dem and rep averaged)
# collapse across dem and rep - find average fam score, then exclude those that you don't need - make index from that.
accuracy_all_two.four <- c_df_political_true %>% group_by(`Item #`, `Image Name`, Par_Combined) %>% filter(measurement == "Fam") %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.four <- accuracy_all_two.four %>% filter(mean < 2.4)
accuracy_all_two.four <- accuracy_all_two.four[[1]]
# accuracy_all_two.four
# length(accuracy_all_two.four)
#given the index, now subset based on the index numbers.
accuracy_all_two.four <- c_df_political_true %>% filter(measurement == "Acc_Abs", `Item #` %in% accuracy_all_two.four) %>% group_by(political_leaning, Par_Combined) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.four
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 64.2 8.33 24 1.70 66.7 45.1 78.6
## 2 Democrat Rep Favoured 55.3 9.37 8 3.31 53.3 46.2 69.8
## 3 Republican Dem Favoured 56.4 8.24 24 1.68 55.7 42.9 78.0
## 4 Republican Rep Favoured 62.2 8.90 8 3.15 65.3 51.4 76.3
#making table
accuracy_all_two.four_table <- accuracy_all_two.four[1:2,-1]
accuracy_all_two.four %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

all items, familiarity thershold < 2.3 (political_true only)
# limited to 2.3
c_df_political_true %>% filter(measurement == "Fam", value < 2.3) %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 6)) + labs(title = "All items, familiarity < 2.3", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

# need to find item numbers of those with Fam < 2.3 (across both dem and rep averaged)
# collapse across dem and rep - find average fam score, then exclude those that you don't need - make index from that.
accuracy_all_two.three <- c_df_political_true %>% group_by(`Item #`, `Image Name`, Par_Combined) %>% filter(measurement == "Fam") %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.three <- accuracy_all_two.three %>% filter(mean < 2.3)
accuracy_all_two.three <- accuracy_all_two.three[[1]]
# accuracy_all_two.three
# length(accuracy_all_two.three)
#given the index, now subset based on the index numbers.
accuracy_all_two.three <- c_df_political_true %>% filter(measurement == "Acc_Abs", `Item #` %in% accuracy_all_two.three) %>% group_by(political_leaning, Par_Combined) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.three
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 65.0 8.51 20 1.90 68.2 45.1 78.6
## 2 Democrat Rep Favoured 55.7 11.2 5 5.03 48 46.9 69.8
## 3 Republican Dem Favoured 55.8 8.39 20 1.88 55.3 42.9 78.0
## 4 Republican Rep Favoured 62.1 10.4 5 4.66 64.9 51.4 76.3
#making table
accuracy_all_two.three_table <- accuracy_all_two.three[1:2,-1]
accuracy_all_two.three %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

all items, familiarity thershold < 2.2 (political_true only)
# limited to 2.2
c_df_political_true %>% filter(measurement == "Fam", value < 2.2) %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 6)) + labs(title = "All items, familiarity < 2.2", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

# need to find item numbers of those with Fam < 2.2 (across both dem and rep averaged)
# collapse across dem and rep - find average fam score, then exclude those that you don't need - make index from that.
accuracy_all_two.two <- c_df_political_true %>% group_by(`Item #`, `Image Name`, Par_Combined) %>% filter(measurement == "Fam") %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.two <- accuracy_all_two.two %>% filter(mean < 2.2)
accuracy_all_two.two <- accuracy_all_two.two[[1]]
# accuracy_all_two.two
# length(accuracy_all_two.two)
#given the index, now subset based on the index numbers.
accuracy_all_two.two <- c_df_political_true %>% filter(measurement == "Acc_Abs", `Item #` %in% accuracy_all_two.two) %>% group_by(political_leaning, Par_Combined) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.two
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 65.2 10.2 12 2.93 69.5 45.1 78.6
## 2 Democrat Rep Favoured 57.7 11.9 4 5.97 57.1 46.9 69.8
## 3 Republican Dem Favoured 53.1 5.55 12 1.60 53.8 42.9 63.0
## 4 Republican Rep Favoured 64.6 10.2 4 5.10 65.3 51.4 76.3
#making table
accuracy_all_two.two_table <- accuracy_all_two.two[1:2,-1]
accuracy_all_two.two %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

Item selection (politically false only)
all items, no familiarity thershold (political_false only)
#all familiarity
c_df_political_false %>% filter(measurement == "Fam") %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 4)) + labs(title = "All items, all familiarity", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

accuracy_all_fam <- c_df_political_false %>% filter(measurement == "Acc_Abs") %>% group_by(political_leaning, Par_Combined) %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_fam
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 55.4 9.94 30 1.81 56.1 39.6 76.1
## 2 Democrat Rep Favoured 64.2 9.71 29 1.80 66.0 46.5 78.3
## 3 Republican Dem Favoured 63.2 9.89 30 1.81 63.2 45.3 78.1
## 4 Republican Rep Favoured 55.0 10.5 29 1.95 56.8 38.2 73.5
accuracy_all_fam %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

all items, familiarity thershold < 3 (political_false only)
# limited to three
c_df_political_false %>% filter(measurement == "Fam", value < 3) %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 4)) + labs(title = "All items, familiarity < 3", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

# need to find item numbers of those with Fam < 3 (across both dem and rep averaged)
# collapse across dem and rep - find average fam score, then exclude those that you don't need - make index from that.
accuracy_all_three_index <- c_df_political_false %>% group_by(`Item #`, `Image Name`, Par_Combined) %>% filter(measurement == "Fam") %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_three_index <- accuracy_all_three_index %>% filter(mean < 3)
accuracy_all_three_index <- accuracy_all_three_index[[1]]
# accuracy_all_three_index
# length(accuracy_all_three_index)
#given the index, now subset based on the index numbers.
accuracy_all_three <- c_df_political_false %>% filter(measurement == "Acc_Abs", `Item #` %in% accuracy_all_three_index) %>% group_by(political_leaning, Par_Combined) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_three
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 55.4 9.94 30 1.81 56.1 39.6 76.1
## 2 Democrat Rep Favoured 64.2 9.71 29 1.80 66.0 46.5 78.3
## 3 Republican Dem Favoured 63.2 9.89 30 1.81 63.2 45.3 78.1
## 4 Republican Rep Favoured 55.0 10.5 29 1.95 56.8 38.2 73.5
#making table
accuracy_all_three_table <- rbind(accuracy_all_three_table, accuracy_all_three[1:2,-1])
accuracy_all_three_table$true_false <- c("Political_True", "Political_True", "Political_False", "Political_False")
accuracy_all_three %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

all items, familiarity thershold < 2.5 (political_false only)
# limited to 2.5
c_df_political_false %>% filter(measurement == "Fam", value < 2.5) %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 5)) + labs(title = "All items, familiarity < 2.5", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

# need to find item numbers of those with Fam < 2.5 (across both dem and rep averaged)
# collapse across dem and rep - find average fam score, then exclude those that you don't need - make index from that.
accuracy_all_two.five <- c_df_political_false %>% group_by(`Item #`, `Image Name`, Par_Combined) %>% filter(measurement == "Fam") %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.five <- accuracy_all_two.five %>% filter(mean < 2.5)
accuracy_all_two.five <- accuracy_all_two.five[[1]]
# accuracy_all_two.five
# length(accuracy_all_two.five)
#given the index, now subset based on the index numbers.
accuracy_all_two.five <- c_df_political_false %>% filter(measurement == "Acc_Abs", `Item #` %in% accuracy_all_two.five) %>% group_by(political_leaning, Par_Combined) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.five
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 55.4 9.94 30 1.81 56.1 39.6 76.1
## 2 Democrat Rep Favoured 64.2 9.71 29 1.80 66.0 46.5 78.3
## 3 Republican Dem Favoured 63.2 9.89 30 1.81 63.2 45.3 78.1
## 4 Republican Rep Favoured 55.0 10.5 29 1.95 56.8 38.2 73.5
#making table
accuracy_all_two.five_table <- rbind(accuracy_all_two.five_table, accuracy_all_two.five[1:2,-1])
accuracy_all_two.five_table$true_false <- c("Political_True", "Political_True", "Political_False", "Political_False")
accuracy_all_two.five %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

all items, familiarity thershold < 2.4 (political_false only)
# limited to 2.4
c_df_political_false %>% filter(measurement == "Fam" , value < 2.4) %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 5)) + labs(title = "All items, familiarity < 2.4", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

# need to find item numbers of those with Fam < 2.4 (across both dem and rep averaged)
# collapse across dem and rep - find average fam score, then exclude those that you don't need - make index from that.
accuracy_all_two.four <- c_df_political_false %>% group_by(`Item #`, `Image Name`, Par_Combined) %>% filter(measurement == "Fam") %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.four <- accuracy_all_two.four %>% filter(mean < 2.4)
accuracy_all_two.four <- accuracy_all_two.four[[1]]
# accuracy_all_two.four
# length(accuracy_all_two.four)
#given the index, now subset based on the index numbers.
accuracy_all_two.four <- c_df_political_false %>% filter(measurement == "Acc_Abs", `Item #` %in% accuracy_all_two.four) %>% group_by(political_leaning, Par_Combined) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.four
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 56.2 9.75 28 1.84 56.5 40.4 76.1
## 2 Democrat Rep Favoured 64.2 9.71 29 1.80 66.0 46.5 78.3
## 3 Republican Dem Favoured 63.5 9.88 28 1.87 63.2 45.3 78.1
## 4 Republican Rep Favoured 55.0 10.5 29 1.95 56.8 38.2 73.5
#making table
accuracy_all_two.four_table <- rbind(accuracy_all_two.four_table, accuracy_all_two.four[1:2,-1])
accuracy_all_two.four_table$true_false <- c("Political_True", "Political_True", "Political_False", "Political_False")
accuracy_all_two.four %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

all items, familiarity thershold < 2.3 (political_false only)
# limited to 2.3
c_df_political_false %>% filter(measurement == "Fam", value < 2.3) %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 6)) + labs(title = "All items, familiarity < 2.3", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

# need to find item numbers of those with Fam < 2.3 (across both dem and rep averaged)
# collapse across dem and rep - find average fam score, then exclude those that you don't need - make index from that.
accuracy_all_two.three <- c_df_political_false %>% group_by(`Item #`, `Image Name`, Par_Combined) %>% filter(measurement == "Fam") %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.three <- accuracy_all_two.three %>% filter(mean < 2.3)
accuracy_all_two.three <- accuracy_all_two.three[[1]]
# accuracy_all_two.three
# length(accuracy_all_two.three)
#given the index, now subset based on the index numbers.
accuracy_all_two.three <- c_df_political_false %>% filter(measurement == "Acc_Abs", `Item #` %in% accuracy_all_two.three) %>% group_by(political_leaning, Par_Combined) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.three
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 57.5 9.53 25 1.91 58.5 40.4 76.1
## 2 Democrat Rep Favoured 64.3 9.88 28 1.87 66.4 46.5 78.3
## 3 Republican Dem Favoured 65.0 9.15 25 1.83 67.6 46.3 78.1
## 4 Republican Rep Favoured 55.5 10.3 28 1.95 57.0 38.2 73.5
#making table
accuracy_all_two.three_table <- rbind(accuracy_all_two.three_table, accuracy_all_two.three[1:2,-1])
accuracy_all_two.three_table$true_false <- c("Political_True", "Political_True", "Political_False", "Political_False")
accuracy_all_two.three %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

all items, familiarity thershold < 2.2 (political_false only)
# limited to 2.2
c_df_political_false %>% filter(measurement == "Fam", value < 2.2) %>%
ggplot() + geom_bar(aes(x = `Image Name`, y = value, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 6)) + labs(title = "All items, familiarity < 2.2", x = "Image Name", y = "Familiarity", fill = "Item Pol. Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

# need to find item numbers of those with Fam < 2.2 (across both dem and rep averaged)
# collapse across dem and rep - find average fam score, then exclude those that you don't need - make index from that.
accuracy_all_two.two <- c_df_political_false %>% group_by(`Item #`, `Image Name`, Par_Combined) %>% filter(measurement == "Fam") %>% summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.two <- accuracy_all_two.two %>% filter(mean < 2.2)
accuracy_all_two.two <- accuracy_all_two.two[[1]]
# accuracy_all_two.two
# length(accuracy_all_two.two)
#given the index, now subset based on the index numbers.
accuracy_all_two.two <- c_df_political_false %>% filter(measurement == "Acc_Abs", `Item #` %in% accuracy_all_two.two) %>% group_by(political_leaning, Par_Combined) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
accuracy_all_two.two
## # A tibble: 4 x 9
## # Groups: political_leaning [2]
## political_leaning Par_Combined mean SD count se median min max
## <chr> <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Democrat Dem Favoured 57.3 9.91 23 2.07 58.5 40.4 76.1
## 2 Democrat Rep Favoured 64.3 10.3 25 2.06 66.7 46.5 78.3
## 3 Republican Dem Favoured 64.7 9.30 23 1.94 67.6 46.3 78.1
## 4 Republican Rep Favoured 55.4 10.4 25 2.08 56.8 38.2 73.5
#making table
accuracy_all_two.two_table <- rbind(accuracy_all_two.two_table, accuracy_all_two.two[1:2,-1])
accuracy_all_two.two_table$true_false <- c("Political_True", "Political_True", "Political_False", "Political_False")
accuracy_all_two.two %>%
ggplot(aes(x = Par_Combined, y = mean, fill = as.factor(political_leaning))) +
geom_bar(stat = 'identity', position = 'dodge') + geom_errorbar(aes(ymin = mean-se, ymax =mean + se), width = .2, position = position_dodge(.9)) + theme_apa(legend.use.title = TRUE) + labs(title ="Accuracy across Dem and Rep favoured items", x = "Item Political Leaning", y = "Percent Accuracy", fill = "Pps Political Leaning") + scale_fill_manual(values=c("#0021F5", "#EA3223"))

Tables of familiarity across True and False items
#fam < 3
accuracy_all_three_table
## # A tibble: 4 x 9
## Par_Combined mean SD count se median min max true_false
## <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl> <chr>
## 1 Dem Favoured 68.5 8.72 44 1.31 69.9 45.1 81.7 Political_True
## 2 Rep Favoured 60.6 11.5 17 2.78 61.0 44.6 77.6 Political_True
## 3 Dem Favoured 55.4 9.94 30 1.81 56.1 39.6 76.1 Political_False
## 4 Rep Favoured 64.2 9.71 29 1.80 66.0 46.5 78.3 Political_False
# accuracy_all_three_table$true_false <- factor(accuracy_all_three_table$true_false, levels = c("Political_True", "Political_False"))
# accuracy_all_three_table %>% ggplot() + geom_bar(aes(x = Par_Combined, y = count, fill = Par_Combined), stat = 'identity', position = 'dodge') + theme_apa(legend.use.title = TRUE) + labs(title = "Item Count, fam < 3", x = "Item Pol. Leaning", y = "Count", fill = "Item Pol. Leaning") + facet_wrap(~true_false) + scale_fill_manual(values=c("#0021F5", "#EA3223"))
#fam < 2.5
accuracy_all_two.five_table
## # A tibble: 4 x 9
## Par_Combined mean SD count se median min max true_false
## <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl> <chr>
## 1 Dem Favoured 66.4 8.89 30 1.62 68.8 45.1 81.7 Political_True
## 2 Rep Favoured 56.9 11.1 11 3.34 58.7 44.6 77.6 Political_True
## 3 Dem Favoured 55.4 9.94 30 1.81 56.1 39.6 76.1 Political_False
## 4 Rep Favoured 64.2 9.71 29 1.80 66.0 46.5 78.3 Political_False
#fam < 2.4
accuracy_all_two.four_table
## # A tibble: 4 x 9
## Par_Combined mean SD count se median min max true_false
## <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl> <chr>
## 1 Dem Favoured 64.2 8.33 24 1.70 66.7 45.1 78.6 Political_True
## 2 Rep Favoured 55.3 9.37 8 3.31 53.3 46.2 69.8 Political_True
## 3 Dem Favoured 56.2 9.75 28 1.84 56.5 40.4 76.1 Political_False
## 4 Rep Favoured 64.2 9.71 29 1.80 66.0 46.5 78.3 Political_False
#fam < 2.3
accuracy_all_two.three_table
## # A tibble: 4 x 9
## Par_Combined mean SD count se median min max true_false
## <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl> <chr>
## 1 Dem Favoured 65.0 8.51 20 1.90 68.2 45.1 78.6 Political_True
## 2 Rep Favoured 55.7 11.2 5 5.03 48 46.9 69.8 Political_True
## 3 Dem Favoured 57.5 9.53 25 1.91 58.5 40.4 76.1 Political_False
## 4 Rep Favoured 64.3 9.88 28 1.87 66.4 46.5 78.3 Political_False
#fam < 2.2
accuracy_all_two.two_table
## # A tibble: 4 x 9
## Par_Combined mean SD count se median min max true_false
## <fct> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl> <chr>
## 1 Dem Favoured 65.2 10.2 12 2.93 69.5 45.1 78.6 Political_True
## 2 Rep Favoured 57.7 11.9 4 5.97 57.1 46.9 69.8 Political_True
## 3 Dem Favoured 57.3 9.91 23 2.07 58.5 40.4 76.1 Political_False
## 4 Rep Favoured 64.3 10.3 25 2.06 66.7 46.5 78.3 Political_False
# final selection
# View(c_df_political_true %>% filter(measurement == "Fam", Par_Combined == "Dem Favoured", political_leaning == "Democrat"))
tmp <- c_df_political_true %>% filter(measurement == "Fam", Par_Combined == "Dem Favoured", political_leaning == "Democrat")
tmp <- rbind(tmp, c_df_political_true %>% filter(measurement == "Fam", Par_Combined == "Rep Favoured", political_leaning == "Democrat"))
tmp <- rbind(tmp, c_df_political_false %>% filter(measurement == "Fam", Par_Combined == "Dem Favoured", political_leaning == "Democrat"))
tmp <- rbind(tmp, c_df_political_false %>% filter(measurement == "Fam", Par_Combined == "Rep Favoured", political_leaning == "Democrat"))
write_xlsx(
tmp,
path = "item_selection - review.xlsx",
col_names = TRUE)
Partisian
# removing unnecessary columns
df_slim_leaning <- df[, c("Item #", "Category", "Image Name", "Headline Summary", "Par_Dem", "Par_Rep", "Par_Combined")]
#add whether the items are democratic favoured or republican favoured
df_slim_leaning$Par_Combined_Categ <- ifelse(df_slim_leaning$Par_Combined > 3.5, "Rep Favoured", "Dem Favoured")
# long format
df_slim_leaning_long <- gather(df_slim_leaning, key = "measurement", value = "value", -c("Item #", "Category", "Image Name", "Headline Summary", "Par_Combined_Categ"))
# adding political variable
df_slim_leaning_long$political_leaning <- "Democrat"
df_slim_leaning_long$political_leaning <- ifelse(str_detect(df_slim_leaning_long$measurement, "Rep") == TRUE, df_slim_leaning_long$political_leaning <- "Republican", df_slim_leaning_long$political_leaning <- "Democrat")
df_slim_leaning_long$political_leaning[str_detect(df_slim_leaning_long$measurement, "Combined")] <- "Combined"
df_slim_leaning_long$Par_Combined_Categ[str_detect(df_slim_leaning_long$measurement, "Combined")] <- "Combined"
df_slim_leaning_long$measurement <- str_replace(df_slim_leaning_long$measurement, c("_Rep"), "")
df_slim_leaning_long$measurement <- str_replace(df_slim_leaning_long$measurement, c("_Dem"), "")
df_slim_leaning_long$measurement <- str_replace(df_slim_leaning_long$measurement, c("_Combined"), "")
#adding classesf
df_slim_leaning_long$Par_Combined_Categ <- factor(df_slim_leaning_long$Par_Combined_Categ, levels = c("Dem Favoured", "Rep Favoured", "Combined"))
df_slim_leaning_long <- df_slim_leaning_long %>% filter(Category == "Political_True" | Category == "Political_Fake")
# removing those high or low in accuracy
# after identifying which items to remove now creating new corrected dfs
df_slim_leaning_long <- df_slim_leaning_long %>% filter(`Item #` %notin% index_remove_all)
df_slim_leaning_long %>% group_by(Category, political_leaning) %>%
summarise(mean = mean(value), SD = sd(value), count = n(), se = (SD/(sqrt(count))), median = median(value), min = min(value), max = max(value))
## # A tibble: 6 x 9
## # Groups: Category [2]
## Category political_leani… mean SD count se median min max
## <chr> <chr> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
## 1 Political_F… Combined 3.44 0.565 59 0.0736 3.34 2.26 4.50
## 2 Political_F… Democrat 3.25 0.549 59 0.0715 3.24 2.12 4.3
## 3 Political_F… Republican 3.63 0.627 59 0.0816 3.6 2.19 4.82
## 4 Political_T… Combined 3.23 0.529 64 0.0661 3.16 2.31 4.50
## 5 Political_T… Democrat 2.99 0.527 64 0.0658 2.90 2.07 4.07
## 6 Political_T… Republican 3.47 0.613 64 0.0766 3.38 2.17 5.15
# Basic histogram partisanship combined
df_slim_leaning_long %>% filter(political_leaning == "Combined") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Partisanship - Combined", x = "Likert Values", y = "Count", subtitle = "If accurate, how favorable to Democrats vs. Republicans (>= 4: Rep)") + geom_vline(aes(xintercept = mean(value)), linetype="dashed") + ylim(0,12.5)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram partisanship combined facet
df_slim_leaning_long %>% filter(political_leaning == "Combined") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Partisanship - Combined", x = "Likert Values", y = "Count", subtitle = "If accurate, how favorable to Democrats vs. Republicans (>= 4: Rep)") + facet_wrap(~Category) + ylim(0,12.5)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram partisanship dems only
df_slim_leaning_long %>% filter(political_leaning == "Democrat") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Partisanship - Democrats Only", x = "Likert Values", y = "Count", subtitle = "If accurate, how favorable to Democrats vs. Republicans (>= 4: Rep)") + geom_vline(aes(xintercept = mean(value)), linetype="dashed") + ylim(0,12.5)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram partisanship dems only facet
df_slim_leaning_long %>% filter(political_leaning == "Democrat") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Partisanship - Democrats Only", x = "Likert Values", y = "Count", subtitle = "If accurate, how favorable to Democrats vs. Republicans (>= 4: Rep)") + facet_wrap(~Category) + ylim(0,12.5)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram partisanship reps
df_slim_leaning_long %>% filter(political_leaning == "Republican") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Partisanship - Republicans Only", x = "Likert Values", y = "Count", subtitle = "If accurate, how favorable to Democrats vs. Republicans (>= 4: Rep)") + geom_vline(aes(xintercept = mean(value)), linetype="dashed") + ylim(0,12.5)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram partisanship reps facet
df_slim_leaning_long %>% filter(political_leaning == "Republican") %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Partisanship - Republicans Only", x = "Likert Values", y = "Count", subtitle = "If accurate, how favorable to Democrats vs. Republicans (>= 4: Rep)") + facet_wrap(~Category) + ylim(0,12.5)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# Basic histogram all partisanship factors
df_slim_leaning_long %>% ggplot(aes(x=value)) + geom_histogram() + theme_apa() + labs(title = "Partisanship - All factors", x = "Likert Values", y = "Count", subtitle = "If accurate, how favorable to Democrats vs. Republicans (>= 4: Rep)") + facet_wrap(~political_leaning)
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
